An enhanced Hill cipher approach for image encryption in steganography

Abstract

In this paper, a new approach is adopted to embed an image within an image using a modified Hill cipher method. The secret image is taken and rearranged to form N 2 × 2 matrices. A key matrix of length 2 × 2 is multiplied with all the N matrices and modulo-256 operation is performed. Then the new N matrices are rearranged to form a one-dimensional matrix. Each element of this matrix is converted to its binary value. A cover image is taken and in each of its pixels, the two least significant bits LSB are replaced with two bits from the binary matrix successively. While retrieving, the last 2 bits from each cover pixel are taken out, arranged to form a binary matrix and are converted into their respective 8-bit values. This one dimensional array is converted into N 2 × 2 matrices. These matrices are multiplied with the inverse of the key matrix to get the original values. These values are then rearranged appropriately to get the secret image which was transmitted.